Expression of the insecticidal protein (IP) genes derived from Bacillus thuringiensis (Bt) in plants has proven extremely difficult. Attempts have been made to express chimeric promoter/Bt IP gene combinations in plants. Typically, only low levels of protein have been obtained in transgenic plants. See, for example, Vaeck et al., Nature 328:33-37, 1987; Barton et al., Plant Physiol. 85:1103-1109, 1987; Fischoff et al., Bio/Technology 5:807-813, 1987.
One postulated explanation for the cause of low expression is that fortuitious transcription processing sites produce aberrant forms of Bt IP mRNA transcript. These aberrantly processed transcripts are non-functional in a plant, in terms of producing an insecticidal protein. Possible processing sites include polyadenylation sites, intron splicing sites, transcriptional termination signals and transport signals. Most genes do not contain sites that will deleteriously affect gene expression in that gene's normal host organism. However, the fortuitous occurrence of such processing sites in a coding region might complicate the expression of that gene in transgenic hosts. For example, the coding region for the Bt insecticidal crystal protein gene derived from Bacillus thuringiensis strain kurstaki (GENBANK BTHKURHD, accession M15271, B. thuringiensis var. kurstaki, HD-1; Geiser et al. Gene 48:109-118 (1986)) as derived directly from Bacillus thuringiensis, might contain sites which prevent this gene from being properly processed in plants.
Further difficulties exist when attempting to express Bacillus thuringiensis protein in an organism such as a plant. It has been discovered that the codon usage of a native Bt IP gene is significantly different from that which is typical of a plant gene. In particular, the codon usage of a native Bt IP gene is very different from that of a maize gene. As a result, the mRNA from this gene may not be efficiently utilized. Codon usage might influence the expression of genes at the level of translation or transcription or mRNA processing. To optimize an insecticidal gene for expression in plants, attempts have been made to alter the gene to resemble, as much as possible, genes naturally contained within the host plant to be transformed.
Adang et al., EP 0359472 (1990), relates to a synthetic Bacillus thuringiensis tenebrionis (Btt) gene which is 85% homologous to the native Btt gene and which is designed to have an A+T content approximating that found in plants in general. Table 1 of Adang et al. show the codon sequence of a synthetic Btt gene which was made to resemble more closely the normal codon distribution of dicot genes. Adang et al. state that a synthetic gene coding for IP can be optimized for enhanced expression in monocot plants through similar methods, presenting the frequency of codon usage of highly expressed monocot proteins in Table 1. At page 9, Adang et al. state that the synthetic Btt gene is designed to have an A+T content of 55% (and, by implication, a G+C content of 45%). At page 20, Adang et al. disclose that the synthetic gene is designed by altering individual amino acid codons in the native Bt gene to reflect the overall distribution of codons preferred by dicot genes for each amino acid within the coding region of the gene. Adang et al. further state that only some of the native Btt gene codons will be replaced by the most preferred plant codon for each amino acid, such that the overall distribution of codons used in dicot proteins is preserved.
Fischhoff et al., EP 0 385 962 (1990), relates to plant genes encoding the crystal protein toxin of Bacillus thuringiensis. At table V, Fischhoff et al. disclose percent usages for codons for each amino acid. At page 8, Fischoff et al. suggest modifying a native Bt gene by removal of putative polyadenylation signals and ATTTA sequences. Fischoff et al. further suggest scanning the native Bt gene sequence for regions with greater than four consecutive adenine or thymine nucleotides to identify putative plant polyadenylation signals. Fischoff et al. state that the nucleotide sequence should be altered if more than one putative polyadenylation signal is identified within ten nucleotides of each other. At page 9, Fischoff et al. state that efforts should be made to select codons to preferably adjust the G+C content to about 50%.
Perlak et al., PNAS USA, 7164 88:3324-3328 (1991), relates to modified coding sequences of the Bacillus thuringiensis cryIA(b) gene, similar to those shown in Fischoff et al. As shown in table 1 at page 3325, the partially modified cryIA(b) gene of Perlak et al. is approximately 96% homologous to the native cryIA(b) gene (1681 of 1743 nucleotides), with a G+C content of 41%, number of plant polyadenylation signal sequences (PPSS) reduced from 18 to 7 and number of ATTTA sequences reduced from 13 to 7. The fully modified cryIA(b) gene of Perlak et al. is disclosed to be fully synthetic (page 3325, column 1). This gene is approximately 79% homologous to the native cryIA(b) gene (1455 of 1845 nucleotides), with a G+C content of 49%, number of plant polyadenylation signal sequences (PPSS) reduced to 1 and all ATTTA sequences removed.
Barton et al., EP 0431 829 (1991), relates to the expression of insecticidal toxins in plants. At column 10, Barton et al. describe the construction of a synthetic AaIT insect toxin gene encoding a scorpion toxin using the most preferred codon for each amino acid according to the chart shown in FIG. 1 of the document.